No abstract

This study is the first to examine the benthic impacts of the otter-trawl fishery on hake in the southern Benguela upwelling region. Infauna were sampled at 4 sites, from southern Namibia to Cape Town by means of 5 replicate grab samples at each of 2 trawling treatments (heavily and lightly trawled areas), paired at each site. The large invertebrate epifauna was also sampled at 2 of these sites using a fine-meshed otter trawl. Sites ranged in depth from 350 to 450 m. Environmental attributes (sediment particle size, total organic carbon, depth, salinity, temperature and dissolved O 2 concentration) were examined along with faunal assemblage composition. Vertical profiles of water mass characteristics showed little long-shore variation, apart from slightly lower O 2 concentrations in the north. Difficulties of pseudo-replication in benthic impact studies are discussed, and methods for circumventing these suggested. There were significant differences in sediment characteristics among the 4 sites, but only 2 sites showed different sediment characteristics between trawling treatments. Studies of species richness, evenness and numbers of infaunal individuals showed little difference between trawling treatments at 3 sites and species diversity was similar between treatments at all 4 sites. Multivariate analyses show marked differences in both infaunal and epifaunal assemblages among the sites and between trawling treatments at all sites. The analyses suggest that differences in trawling intensity are at least partially responsible for significant variation in benthic assemblage composition between heavily and lightly trawled areas. These findings contrast to those in shallower waters in the northern hemisphere, where infauna are more sensitive to trawling than epifauna. This study shows that epifaunal abundances, number of species and species diversity decrease with increasing trawling intensity, and that there are also considerable changes in epifaunal assemblages in more heavily trawled sites.

No abstract

Using long‐term survey data, changes in demersal faunal communities in the Benguela Current Large Marine Ecosystem were analysed at community and population levels to provide a comparative overview of the occurrence and timing of regime shifts. For South Africa, the timing of a community‐level shift observed in the early 1990s, and of a lesser shift observed in the mid‐2000s, corresponded well with the results of other studies that showed environmental, community‐level or population‐level changes at similar times, suggesting that environmental forcing had played a role. Several population‐level shifts were detected for Namibia; these and a regime shift in the overall community identified for this country corresponded well to the timing of severe environmental perturbations and an extensive regime shift in the pelagic ecosystem of this area. However, the interpretation of these shifts was confounded by changes in sampling gear; closer scrutiny of the types of species affected and the direction of shifts (increase/decrease) in relation to the timing and nature of sampling gear modifications, revealed that the observed shifts were potentially an artefact of gear changes. This highlighted the importance of accounting for changes in sampling protocols during the analysis and interpretation of long‐term data. For Angola, a community level shift in the mid‐2000s and population‐level changes for a few species (mainly positive), could not have been influenced by gear changes which took place mainly before the onset of the time series under consideration. However, no clear environmental or anthropogenic changes that could have influenced these shifts were obvious.